When we see we synthesize the activity of millions of photoreceptors to form a perceptual image of a visual scene. Our long-term goal is to understand this process, something that requires knowing both the topography of the receptor mosaic and how the activity of the individual cones within it contributes to perception. We approach the broad problem of how information is extracted from the retinal image with a specific focus on the mechanisms of color perception. We propose a series of experiments using a novel combination of imaging, psychophysics, and modeling in the same living human subjects to determine how individual L, M, and S cones are arranged across the retina and the specific contributions they make to color appearance. In Specific Aim 1 we will use adaptive optics imaging to characterize the topography of the adult human trichromatic mosaic. We will test whether L and M cones are arranged randomly or non-randomly across the retina. This information is key to understanding the developmental and genetic factors responsible for the organization of the trichromatic mosaic as well as the neural circuitry and visual processes underlying spatial and color vision. In Specific Aim 2 we will use adaptive optics stimulus delivery with concurrent retinal tracking to uncover the contributions of individual L, M, and S cones to color appearance. This approach allows a direct link between the activity of individual cones and visual perception, and as such provides the aggregate action of the entire neural visual system uncontaminated by optical or sampling effects. We will test the specific hypothesis that the contributions of individual cones to appearance depend on the local retinal mosaic and not just on photopigment. We predict that cones entirely within cones of like-type will not contribute to color. The results of these experiments will constrain models of spatial and color processing and their possible neural substrates. In Specific Aim 3 will assess the extent to which the visual system optimally extracts color information from the retinal image. By comparing the subjective appearance of stimuli that excite particular cones with predictions generated by ideal-observer appearance models we will test the hypothesis that cone contributions to color appearance are tailored within an individual retina by a mechanism that seeks to maximize perceptual veridicality. To the extent that this model accounts for the data, the principles it embodies may be expected to generalize to other aspects of visual processing. Overall these experiments provide an unprecedented opportunity to determine the individual cone inputs to the mechanisms of human color vision, and to delineate the principles that guide them. The knowledge we will gain will also allow a new understanding of the factors involved in cone detection, and better and more accurate models of both the mechanisms underlying visual perception and the genetic and developmental processes that come into play during retinal development. PUBLIC HEALTH RELEVANCE: This grant will positively impact public health because a clear understanding of retinal topography and the early stages of normal vision is needed to fully understand the effects of abnormalities in retinal organization, or photoreceptor loss, on vision. For example, this work will be important for understanding the potential benefits of retinal implants and strategies to restore photoreceptor function in disease, as well any future attempts to cure color blindness with gene therapies that allow photoreceptors to express new pigments. Finally, the imaging and sensitivity mapping tools we will develop may eventually permit earlier diagnosis of retinal disease and better monitoring of disease progression, and may also allow us to probe the earliest disease-related changes in the structure and function of individual receptors.